## MIDAS MAR-1 test
from midas import Midas
from sklearn.preprocessing import MinMaxScaler
import numpy as np
import pandas as pd
 
directory = 'mar1/data_tmp/'

B = 100

for b in range(1,101):
	print("Iteration: " + str(b))
	file = 'mar1_draw_' + str(b) + '.csv'
	data = pd.read_csv(directory+file)

	imputer = Midas(vae_layer= False, seed=89)
	
	imputer.build_model(data)

	imputer.train_model(training_epochs = 5)

	imputer.batch_generate_samples(m=10)

	imp = 0
	for dataset in imputer.output_list:
		dataset.to_csv(directory + 'midas_'+str(b)+'_'+str(imp+1)+'.csv', index=False)
		imp += 1

	del imputer
